Re: linux-next: Signed-off-by missing for commits in the arm-soc tree

[Florian Fainelli]

On 07/16/2018 02:57 AM, Russell King - ARM Linux wrote:
> On Mon, Jul 16, 2018 at 05:24:08PM +0800, Chen-Yu Tsai wrote:
>> On Mon, Jul 16, 2018 at 5:13 PM, Florian Fainelli <f.fainelli@xxxxxxxxx> wrote:
>>>
>>>
>>> On 07/15/2018 03:50 PM, Olof Johansson wrote:
>>>> Thanks Stephen, I keep saying every time you catch these that I need
>>>> to run the same script. :(
>>>>
>>>> Florian, I wonder if this happened when you rebased to squash in the fix?
>>>
>>> Humm, could be, all I did (and it was not the first time) was to do an
>>> interactive rebase with --preserve-merges. AFAICT, all of these commits
>>> came from Eric's pull request, so what I typically do is just sign off
>>> on the merge commit, but do not apply my SoB to all commits coming from
>>> that pull request if that makes sense?
>>
>> When you rebase, if the commit is re-applied, the committer changes to
>> you, so you need to add your SoB to all commits you rebased.
>>
>> I suppose you could override it with GIT_COMMITTER_NAME and
>> GIT_COMMITTER_EMAIL environment variables, but that sort of distorts
>> the SoB if any changes were added due to rebasing, such as resolution
>> of merge conflicts.
> 
> A different point: should *anyone* be rebasing commits that they did not
> themselves commit.  Discuss.
> 
> IMHO no - the commits were made public, and anyone could pull the tree
> from which they came in order to do further work before submitting that.
> Rebasing changes the commit IDs, which can lead to duplicate commits
> ending up in mainline.  With other changes on top, this causes totally
> unnecessary conflicts.
> 

Well in that case, I thought it would be cleaner to just squash Arnd's
build fix since the commit had not been merged into arm-soc, but I will
keep in mind not to do that anymore.
-- 
Florian